<< 10-09-2013 >>
00:08:05reactormonknope.
00:19:15*reactormonk quit (Ping timeout: 260 seconds)
00:43:10*q66 quit (Quit: Leaving)
00:50:35*Yeri quit (Quit: Yeri)
01:04:13*ltbarcly joined #nimrod
01:14:18*shodan45 quit (Ping timeout: 245 seconds)
01:19:00*io2 quit ()
01:21:50*brson quit (Ping timeout: 264 seconds)
01:23:15*DAddYE quit (Remote host closed the connection)
01:23:51*DAddYE joined #nimrod
01:27:03*ltbarcly quit (Ping timeout: 276 seconds)
01:29:00*DAddYE quit (Ping timeout: 276 seconds)
01:29:15*ltbarcly joined #nimrod
01:33:46*ltbarcly quit (Ping timeout: 245 seconds)
01:35:42*ltbarcly joined #nimrod
01:36:36*jpoirier joined #nimrod
01:40:14*ltbarcly quit (Ping timeout: 240 seconds)
01:42:14*ltbarcly joined #nimrod
01:46:41*ltbarcly quit (Ping timeout: 245 seconds)
01:47:46*ltbarcly joined #nimrod
01:48:16*Yeri joined #nimrod
01:54:36*ltbarcly quit (Ping timeout: 256 seconds)
01:56:44*ltbarcly joined #nimrod
02:01:03*ltbarcly quit (Ping timeout: 240 seconds)
02:03:17*ltbarcly joined #nimrod
02:08:08*ltbarcly quit (Ping timeout: 260 seconds)
02:10:16*ltbarcly joined #nimrod
02:14:03*ltbarcly quit (Client Quit)
02:29:27*ltbarcly joined #nimrod
02:34:15*ltbarcly quit (Ping timeout: 256 seconds)
02:34:38*DAddYE joined #nimrod
02:36:26*ltbarcly joined #nimrod
02:37:25*DAddYE_ joined #nimrod
02:39:02*DAddYE quit (Ping timeout: 240 seconds)
02:41:29*ltbarcly quit (Ping timeout: 268 seconds)
02:43:25*ltbarcly joined #nimrod
02:47:50*ltbarcly quit (Ping timeout: 240 seconds)
02:49:57*ltbarcly joined #nimrod
02:54:39*ltbarcly quit (Ping timeout: 256 seconds)
02:55:54*ltbarcly joined #nimrod
03:01:13*ltbarcly quit (Ping timeout: 268 seconds)
03:02:25*ltbarcly joined #nimrod
03:07:07*ltbarcly quit (Ping timeout: 256 seconds)
03:08:55*ltbarcly joined #nimrod
03:14:10*ltbarcly quit (Ping timeout: 268 seconds)
03:15:56*ltbarcly joined #nimrod
03:20:14*ltbarcly quit (Ping timeout: 240 seconds)
03:23:01*ltbarcly joined #nimrod
03:30:21*ltbarcly quit (Ping timeout: 256 seconds)
03:32:31*ltbarcly joined #nimrod
03:36:32*Yeri quit (Quit: Yeri)
03:37:38*ltbarcly quit (Quit: Computer has gone to sleep.)
03:48:42*ltbarcly joined #nimrod
03:53:19*ltbarcly quit (Ping timeout: 246 seconds)
03:54:42*ltbarcly joined #nimrod
03:59:02*ltbarcly quit (Ping timeout: 240 seconds)
04:00:11*ltbarcly joined #nimrod
04:04:36*ltbarcly quit (Ping timeout: 245 seconds)
04:06:40*ltbarcly joined #nimrod
04:11:11*ltbarcly quit (Ping timeout: 246 seconds)
04:12:38*ltbarcly joined #nimrod
04:17:02*ltbarcly quit (Ping timeout: 240 seconds)
04:19:06*ltbarcly joined #nimrod
04:21:34*ltbarcly quit (Client Quit)
04:34:07*OrionPK quit (Read error: Connection reset by peer)
05:05:17*ltbarcly joined #nimrod
05:07:23*ltbarcly quit (Client Quit)
05:36:58*ltbarcly joined #nimrod
05:41:49*ltbarcly quit (Ping timeout: 256 seconds)
05:42:26*ltbarcly joined #nimrod
05:47:43*ltbarcly quit (Ping timeout: 268 seconds)
05:48:57*ltbarcly joined #nimrod
05:58:15*ltbarcly quit (Ping timeout: 256 seconds)
06:00:27*ltbarcly joined #nimrod
06:05:36*ltbarcly quit (Ping timeout: 268 seconds)
06:06:28*ltbarcly joined #nimrod
06:11:52*nebiros quit (Quit: ZNC - http://znc.in)
06:11:58*d34th quit (Quit: Deleted System32)
06:12:16*d34th joined #nimrod
06:13:33*ltbarcly quit (Ping timeout: 256 seconds)
06:14:22*nebiros joined #nimrod
06:14:26*nebiros is now known as Guest44283
06:14:57*ltbarcly joined #nimrod
06:44:27*ltbarcly quit (Ping timeout: 268 seconds)
06:45:28*ltbarcly joined #nimrod
06:54:55*ltbarcly quit (Ping timeout: 256 seconds)
06:56:23*zahary quit (Ping timeout: 246 seconds)
06:56:27*ltbarcly joined #nimrod
07:04:11*ltbarcly quit (Ping timeout: 268 seconds)
07:05:28*ltbarcly joined #nimrod
07:14:45*ltbarcly quit (Ping timeout: 256 seconds)
07:15:57*ltbarcly joined #nimrod
07:15:58*silven joined #nimrod
07:20:50*ltbarcly quit (Ping timeout: 268 seconds)
07:21:27*ltbarcly joined #nimrod
07:26:05*ltbarcly quit (Ping timeout: 256 seconds)
07:27:57*ltbarcly joined #nimrod
07:28:23*DAddYE_ quit (Remote host closed the connection)
07:29:01*DAddYE joined #nimrod
07:30:39*io2 joined #nimrod
07:33:13*DAddYE quit (Ping timeout: 240 seconds)
07:35:38*ltbarcly quit (Ping timeout: 268 seconds)
07:36:56*ltbarcly joined #nimrod
07:41:57*ltbarcly quit (Ping timeout: 256 seconds)
07:42:57*ltbarcly joined #nimrod
07:50:26*ltbarcly quit (Ping timeout: 268 seconds)
07:51:25*ltbarcly joined #nimrod
07:59:04*Araq_ joined #nimrod
08:01:26*ltbarcly quit (Ping timeout: 264 seconds)
08:02:28*silven quit (Remote host closed the connection)
08:03:27*ltbarcly joined #nimrod
08:04:45*q66 joined #nimrod
08:07:50*ltbarcly quit (Ping timeout: 240 seconds)
08:10:00*ltbarcly joined #nimrod
08:16:27*faassen joined #nimrod
08:18:46*ltbarcly quit (Ping timeout: 245 seconds)
08:19:47*Araq_ quit (Quit: ChatZilla 0.9.90.1 [Firefox 23.0.1/20130814063812])
08:20:01*ltbarcly joined #nimrod
08:24:50*ltbarcly quit (Ping timeout: 264 seconds)
08:27:02*ltbarcly joined #nimrod
08:31:26*ltbarcly quit (Ping timeout: 240 seconds)
08:33:06*ltbarcly joined #nimrod
08:37:05*ltbarcly quit (Client Quit)
08:50:12*jpoirier quit (Quit: Leaving...)
08:54:02zahary_Araq, it was mentioned in my original commit here:
08:54:03zahary_https://github.com/Araq/Nimrod/commit/cb6507759e7c8496b16a4250946d089491e3e7eb
08:54:42zahary_it basically produces faster compiler, it's recommended for the caas mode anyway and fixes the Aporia build on a mac - so right now, it seems like a better choice
08:54:50zahary_... for the compiler itself
09:07:22*faassen quit (Ping timeout: 245 seconds)
09:07:23*faassen joined #nimrod
09:11:48*Guest44283 quit (*.net *.split)
09:11:48*d34th quit (*.net *.split)
09:11:48*BitPuffin quit (*.net *.split)
09:18:28*io2 quit (*.net *.split)
09:18:28*Associat0r quit (*.net *.split)
09:18:28*XAMPP quit (*.net *.split)
09:18:29*zahary_ quit (*.net *.split)
09:18:29*companion_cube quit (*.net *.split)
09:18:29*profmakx quit (*.net *.split)
09:18:29*mal`` quit (*.net *.split)
09:18:30*ponce quit (*.net *.split)
09:18:32*photex quit (*.net *.split)
09:18:32*Roin quit (*.net *.split)
09:18:32*q66 quit (*.net *.split)
09:18:32*Zor quit (*.net *.split)
09:18:33*JStoker quit (*.net *.split)
09:18:33*Amrykid quit (*.net *.split)
09:18:35*tumak quit (*.net *.split)
09:18:35*rndbit quit (*.net *.split)
09:18:35*comex quit (*.net *.split)
09:18:36*vegai quit (*.net *.split)
09:18:36*Trixar_za quit (*.net *.split)
09:18:36*Araq quit (*.net *.split)
09:18:36*eco quit (*.net *.split)
09:18:36*dom96 quit (*.net *.split)
09:18:36*Reisen quit (*.net *.split)
09:21:08*BitPuffin joined #nimrod
09:21:08*d34th joined #nimrod
09:21:08*Guest44283 joined #nimrod
09:21:08*q66 joined #nimrod
09:21:08*io2 joined #nimrod
09:21:08*Associat0r joined #nimrod
09:21:08*profmakx joined #nimrod
09:21:08*zahary_ joined #nimrod
09:21:08*XAMPP joined #nimrod
09:21:08*Zor joined #nimrod
09:21:08*JStoker joined #nimrod
09:21:08*mal`` joined #nimrod
09:21:08*companion_cube joined #nimrod
09:21:08*rndbit joined #nimrod
09:21:08*ponce joined #nimrod
09:21:08*eco joined #nimrod
09:21:08*comex joined #nimrod
09:21:08*Amrykid joined #nimrod
09:21:08*Trixar_za joined #nimrod
09:21:08*Reisen joined #nimrod
09:21:08*dom96 joined #nimrod
09:21:08*Araq joined #nimrod
09:21:08*tumak joined #nimrod
09:21:08*Roin joined #nimrod
09:21:08*photex joined #nimrod
09:21:08*vegai joined #nimrod
09:23:07*silven joined #nimrod
09:25:49rndbitis anyone using latest aporia (master, last commit on Aug 30, 2013)? not sure if its broken in repo or its my problem
09:28:21rndbitactually nvm, recompiling with optimizations made it work
09:36:31*Associat0r quit (Read error: Operation timed out)
09:37:13*Associat0r joined #nimrod
09:37:13*Associat0r quit (Changing host)
09:37:13*Associat0r joined #nimrod
10:11:37*BitPuffin_ joined #nimrod
10:15:02*BitPuffin quit (Ping timeout: 240 seconds)
10:55:55*Associ8or joined #nimrod
10:55:55*Associ8or quit (Changing host)
10:55:55*Associ8or joined #nimrod
10:58:40*Associat0r quit (Ping timeout: 264 seconds)
11:53:36*Araq_ joined #nimrod
11:55:51*Araq_ quit (Client Quit)
12:07:03*Associ8or quit (Quit: Associ8or)
12:47:33*io2 quit ()
12:59:55*Guest44283 is now known as nebiros
12:59:56*nebiros quit (Changing host)
12:59:56*nebiros joined #nimrod
13:07:50*faassen left #nimrod (#nimrod)
13:54:55*JStoker quit (Quit: JStoker is gone :()
13:59:31*JStoker joined #nimrod
14:37:02*silven quit (Quit: No Ping reply in 180 seconds.)
14:37:21*BitPuffin_ quit (Remote host closed the connection)
14:37:24*silven joined #nimrod
15:10:47dom96rndbit: What's the problem?
15:29:49*Mat2 joined #nimrod
15:29:54Mat2hi @ all
16:02:19*ltbarcly joined #nimrod
16:02:35*ltbarcly quit (Client Quit)
16:03:09*ltbarcly joined #nimrod
16:04:26*ltbarcly_ joined #nimrod
16:06:14rndbiti noticed strange thing, following example plrints 1. why would it print 1 for uninitialized variable? i suppose one would expect it to be 0. http://paste2box.com/s/#/file/ThfdGwQ-FjBqpXiumaTC6QNEcjA/sVMev9Lz1o.txt/?p=E4otjy2VBf1ILkQEFBnGy4M8ye4X
16:07:30*ltbarcly quit (Ping timeout: 245 seconds)
16:08:44Mat2hi rndbit
16:09:16rndbithey
16:11:24Mat2why is your variable not public declared ?
16:13:50rndbitprobably because im just experimenting and i dont quite know how things work
16:16:30Mat2first change:
16:16:49Mat2var a = TSomeObj()
16:16:56Mat2var a: TSomeObj
16:17:09Mat2<- you defined an object as type
16:17:35Mat2then
16:17:51Mat2echo ("Variable: ",a.Variable)
16:18:22Mat2this compiles here and executing the executable return: 0, as expected
16:19:32rndbiti thought () implied construction of an object where lack of () wiykd gave defined object as type
16:19:59Mat2look here: https://gist.github.com/anonymous/6511834
16:20:46*shodan45 joined #nimrod
16:20:47rndbitah i see.. var a: TSomeObj is like declaring object on stack without manually calling ctor
16:21:44Mat2I do not know what you mean with ctor but yes, type declarations are reserved on the heap as I know
16:23:08rndbitctor = constructor
16:23:35rndbiti thought it would be equivalent of c++ TSomeObj a;
16:24:55Mat2please note I do not have any experience with object allocation in C++, but typical in C objects would be defined as structures and also allocated on the heap
16:25:29rndbitheap or stack, programmer's choice. anyways..
16:26:09rndbitscanning through docs i could not find what would look like familiar class. so object methods are always defined outside of object (type) body with first param as reference to that object?
16:27:51Mat2(only in languages as Forth it would be usable to allocate objects on a stack, despite that even in this case objects are more licle declared as dictionary entrys extending the compiler)
16:28:20Mat2^entries
16:29:20Mat2Nimrod uses another appoach to objects as C++, you will not find an equivalence of its class concept
16:31:08rndbitgotcha. as i see types are more like c structs
16:31:37Mat2yes, you would define objects in C in a similar way
16:32:13zahary_Mat2, I think you are spreading misconceptions here
16:32:30Mat2feel free to correct my view
16:32:42zahary_regular values (objects, tuples, etc) will be allocated on the stack just like in C++
16:33:11*silven quit (Remote host closed the connection)
16:33:12zahary_about the original question, I actually think that the compiler missed to produce a compile-time error here
16:33:37zahary_the object initialization syntax var x = TFoo(...) requires that you specify all fields
16:34:19zahary_this is clearly not the case with your code and I would submit it as bug - why the value of 1 is printed is even more peculiar, I would look at the generated C code to see what's happening
16:34:34Mat2ok, I think to have another viewpoint of what a stack is (that's because of my Forth background)
16:35:17zahary_http://en.wikipedia.org/wiki/Call_stack usually as defined by the CPU architecture
16:35:41rndbiterr call stack and stack are two different things
16:35:42Mat2well, that explains the misconception
16:35:58Mat2a call stack is quite different from a data-stack
16:36:10zahary_sure
16:36:59Mat2just replace heap in my answers with call-stack
16:37:24rndbitand then it sounds absolutely weird :D
16:37:31rndbitdata is not allocated on callstack xD
16:37:44Mat2that's right *lol*
16:37:53rndbithttp://en.wikibooks.org/wiki/X86_Disassembly/The_Stack \o/
16:38:48Mat2I know what a stack is. Its only Forth distinguish a data-stack, a return-stack and a list (called vocabulary)
16:39:14rndbitah.. hard to tell when people are fooling around without seeing their faces :)
16:40:29Mat2the return stack and vocabulary are allocated on the heap in most C implementations (that refuses things for me)
16:41:38Mat2by the way, I wounder why objects are allocated on the call-stack ?
16:41:48Mat2^wonder
16:42:03Mat2anyhow
16:42:14*DAddYE joined #nimrod
16:42:23Mat2hi DAddYE
16:42:30DAddYEhi Mat2!
16:44:09*silven joined #nimrod
16:45:25zahary_Mat2, allocating on the call-stack is extremely efficient - when a function is entered, only the stack top register must be modified once to allocate memory for all the local variables in the function
16:49:47Mat2well the seperation of data and addresses have one advantage: Because most cpu's feature a large register set, a compiler is able to transform accesses to stack entries into direct register assignements and in addition the hardware-stack is free for holding addresses only - resulting in both lesser BTB mispredictions and avoidance of memory accesses
16:51:42Mat2(this optimization is not possible in languages as C however)
16:52:14Mat2so I think the Nimrod compiler has no choice here
16:52:49Mat2and in this case, yes - allocating data on the call-stack is more efficient
16:54:15zahary_where is data allocated in the other scheme? in a shadow stack that roughly follows the expansion and contraction of the call-stack or directly on the heap using some allocation algorithm for long-lived objects?
16:57:56Mat2an adaptive compiler is able to map a set of registers to a stack (may it be a data or call-stack does not matter). Because the object itself is static, all references can be transformed into direkt register assignements, simple
16:58:00*brson joined #nimrod
16:59:55zahary_well, the C compilers do this all the time (elect local variables to be stored in register locations), but they surely run out of registers frequently
17:00:08Mat2but as written, that's not possible in static compiled languages as C, because the whole runtime state is not known at compile time, whereby in case of a dynamic compiler (JIT compiler for example)
17:00:40Mat2such environment is able do support register dynamic sheduling
17:01:02zahary_what is this required "runtime state" that must be known?
17:02:09Mat2all dependencies and values for all referenced variables in your example
17:02:48Mat2the compiler get this information though an interpreter and compilation can use it of course
17:03:25zahary_dependencies? like other call-stack locations? that reside at fixed offsets form the current stack base pointer?
17:04:49Mat2yes and in addition the dynamic state of used variables
17:05:11Mat2I give you an example:
17:05:20Mat2You define an object in Nimrod
17:05:44Mat2this object has an variable
17:06:10Mat2somewhere in the source code the variable content is changed, overwritten whatever
17:07:00Mat2because execution is performed though interpretation, all these changes to the variable (the dynamic state at a given execution time) can be tracked
17:07:36Mat2a JIT compiler is then able to use this tracked information for optimisations as earlier described
17:07:59Mat2as long as the object itself is static and do not change
17:09:02Mat2this lead to compiled code which only references internal registers (in most cases)
17:09:19Mat2you can't aplly this kind of optimization with a static compiler
17:09:51zahary_can you describe one particular optimization? are we talking about using guards like "if objects.wasnt_modified? : enter known fast path else: reinterpret and store knowledge of current object state
17:10:14Mat2ok, I try it:
17:10:27zahary_basically, stuff like PIC, etc
17:11:32zahary_feel free to send me a paper about it if don't feel like describing it
17:12:47Mat2http://www.cl.cam.ac.uk/research/srg/netos/vee_2012/papers/p63.pdf
17:13:33zahary_thanks
17:14:22Mat2(please note, the approach decribed in this paper uses a register based VM where I would prefer a stack based one. Additionally there register allocation scheme is staic, I tried to explain a dynamic approach to you)
17:15:23Mat2^static
17:15:56Mat2for a quick look, read 3.3 Register Mapping Translator
17:16:07Mat2following
17:19:02Mat2I think you do not speak German ?
17:19:10zahary_nope
17:20:38Mat2this topic is somewhat complex and I fear my English is not good enough for describing dynamic register-transformation optimizations in detail, sorry
17:21:28Mat2anyway, we are way off topic
17:27:54Mat2and I think its correct to say, an object in Nimrod resembles more of a structure in C (if this object is allocated on a stack (sigh!) or the heap does not matter in tis case)
17:29:22Mat2as an object structure in C++ which hold a vtable of method references bound to it (hope this is the right terminology)
17:31:27zahary_you mean "than an object structure in C++" (i.e. stressing that nimrod doesn't use vtables?)
17:31:46zahary_otherwise, I've read only part of the paper, but the C compilers certainly map stack locations to registers with similar register allocation algorithms - you can see this in the generated code
17:36:04Mat2you oversee the point here: The SWIFT compiler *ensures* this optimization though some code-transformations in most situations
17:41:03Mat2my design *grant* it (by using a compact immediate-code representation in CPS form)
17:41:50Mat2also typical such optimizations will be applied at runtime by a JIT compiler
17:42:16Araqwhat the fuck is wrong with our "object of" code generation? it's like the 4th bug about it triggered by a trivial program ...
17:42:38Mat2hi Araq
17:43:36Araqhi Mat2
17:44:23AraqMat2: nobody here understands your "dynamic register allocation" algorithms ... ;-)
17:44:32Mat2I see
17:44:43Araqwe need to see the code once it's ready
17:45:19Araqthis Forth stuff is completely alien technology :D
17:45:48Mat2most of this is known since the 70's
17:46:11Mat2anyhow, its not common, right
17:47:58*BitPuffin joined #nimrod
17:48:42Mat2at least Dalvik's author understand me *sniff* ;)
17:50:37Araqwell first you need to explain why dynamic register allocation is even desirable ;-)
17:52:41Mat2as written, most memory references can be transformed into direct register assignements for the whole runtime lifetime. It is also the preferable for efficient JIT compilation in ressource-restricted environments
17:52:58Mat2^a preferable way
17:53:44Araqso you mean there is potential aliasing that keeps a static compiler to keep the variable in a register but a JIT can do it?
17:53:57Araq*keeps from keeping
17:53:59Mat2yes
17:55:13Mat2it is even possible to completly get rid of stack and heap allocations alltogether
17:55:22rndbitis it possible to have one statement for variable declaration and allocation instead of two like in following example? http://paste2box.com/s/#/file/V9ud8wGJL3QeYkBZ0TG-FiBAnU0/OxG4Wx3Vzz.txt/?p=uI8N3YOK38j2OjLkRn6abuLpfg8X
17:55:52Araqrndbit: var n = PNode(data: 9)
17:56:22rndbitand its same thing right?
17:56:47Araqyes
17:56:57rndbitgotcha, thanks \o/
17:58:17rndbitoh Araq i was thinking about discardable and what you told me yesterday. well if proc declaration changes and it no longer returns value - no matter if we have discardable or not compiler would raise error where proc is used improperly. so im again back at not getting the point. please enlighten me if its not too annoying ^_^
18:01:46Araqwhat if it changes and it now returns an important value?
18:02:40rndbitis not programmer smart enough to use it where it should be used?
18:03:15rndbitusually if there is a case where same action has to be done with return value then its only logical to include that action in proc that returns object in the first place
18:04:24Araqyeah the programmer is not "smart enough"; hence the need for static typing
18:05:46rndbitso there is no good answer then, gotcha
18:06:50*reactormonk joined #nimrod
18:07:18Araqwell you either see the value in static typing or you don't
18:07:32Araqif you see the value, 'discard' is a consequent feature
18:07:55Araqand as I said, in practice it does find bugs just like the rest of static typing
18:08:02reactormonkis seq[TClient] supposed to be initialized by @[] or nil?
18:08:07Araqnil
18:08:15reactormonkno not nill + lazy alloc?
18:08:29rndbiti just want to understand how could it FIND bugs since it would raise error at compile time just the same way as improper use would
18:08:50Araqbut it doesn't, think about it:
18:08:54Araqproc p() = ...
18:09:04Araq# lots of code calls 'p' in a void context
18:09:17Araqnow you change it to 'proc p(): int'
18:09:36Araqand the compiler tells you you might have got it wrong
18:09:56Araqafter all, if you return additional information, that surely is useful?
18:10:37rndbitcould be useful or could be not, its not like return values are ALWAYS used
18:10:53rndbitbesides such compiler behavior implies programmer has no idea wtf he is doing ^_^
18:10:54Araqthat's why we have .discardable.
18:12:26rndbitreminds me microsoft's decision to not implement default parameters in c# because its easy to get things wrong using them
18:13:26AraqC# has default parameters
18:13:57Araqyour argument amounts to "I don't need compiler checking, I know what I'm doing"
18:16:55rndbitc# does not have default parameters in constructors and methods unless they were added in 4.0 or 4.5. and yeah, well more like "i dont need to tell compiler that i do nothing when im not doing anything". anyway not that its hard to modify compiler to implicitly discard everything \o/
18:18:24AraqI think 3.5 added default params but maybe it was 4.0
18:18:42Araqin any case, they work whenever I use them :P
18:19:22rndbitoh yeah they added them in 4.0. bitches finally listened to the voice of reason.
18:20:29Araqthey had to, it's important for COM and VB had it too
18:21:09Araqbtw ever programmed against an API that uses return values for error reporting?
18:21:41Araqis there a slight chance you can see the value in a compiler that makes you handle errors?
18:22:56*Associat0r joined #nimrod
18:23:03*Associat0r quit (Changing host)
18:23:03*Associat0r joined #nimrod
18:23:35rndbityes, and in that case whenever i add an error code i list all references to modified call and modify code appropriately. but i suppose in this case discard can really be useful
18:23:58rndbiterror codes are evil though..
18:29:25Araqit doesn't matter, posix and the winapi use error codes and for technical reasons they won't go anywhere
18:30:29Araq# XXX object initialization? but not necessary for temps, is it?
18:30:44Araqnow that explains the bug ...
18:52:28*reactormonk quit (Ping timeout: 246 seconds)
19:08:31*reactormonk joined #nimrod
19:14:33Mat2hi reactormonk
19:23:52reactormonkjust phasing in & out
19:24:21Araqreactormonk: the JS bug is not due to a missing "import" (what's that in JS anyway?)
19:24:34Araqbut it surely is a bug
19:24:50reactormonkAraq, there is no import, but the code isn't imported into the resulting .js file
19:25:17Araqah
20:04:35NimBotAraq/Nimrod master b618706 Araq [+1 ±4 -0]: fixes #575
20:25:26*shevy left #nimrod ("I'll be back ... maybe")
20:26:16*reactormonk quit (Ping timeout: 245 seconds)
20:37:16Mat2get some sleep, ciao
20:37:30*Mat2 quit (Quit: Verlassend)
21:21:07NimBotAraq/Nimrod master 8f925e0 ventor3000 [+1 ±0 -0]: Added basic2d module... 2 more lines
21:21:07NimBotAraq/Nimrod master e0e4458 ventor3000 [+0 ±1 -0]: Added file header
21:21:07NimBotAraq/Nimrod master 6548cc5 ventor3000 [+0 ±1 -0]: Some minor changes
21:21:07NimBotAraq/Nimrod master e476a1f ventor3000 [+0 ±1 -0]: Fixed stupid mistake when clamping acos
21:21:07NimBot5 more commits.
21:36:43Araqping zahary_
22:09:43*OrionPK joined #nimrod
22:22:23NimBotAraq/Nimrod master 73333eb Araq [+0 ±2 -0]: fixes #588
22:22:23NimBotAraq/Nimrod master 39acf3f Araq [+0 ±1 -0]: fixes #566
22:22:23NimBotAraq/Nimrod master 3f2d954 Araq [+0 ±3 -0]: c2nim: added some scope operator parsing
22:22:23NimBotAraq/Nimrod master 2176a57 Araq [+2 ±0 -0]: Merge branch 'master' of github.com:Araq/Nimrod
22:24:55BitPuffinAraq: Have you started binding Qt on your own?
22:25:09Araqhell no
22:25:18BitPuffinhehe :=
22:25:19BitPuffin:)*
22:26:15BitPuffinMaybe fowl would be an appropriate person to do it?
22:26:18AraqI hope what we have now is useful enough; I need to hunt bugs and finish my VM instead
22:26:30Araqfowl is pretty busy these days ...
22:26:42BitPuffinunfortunately me too :/
22:27:17BitPuffinMaybe you've noticed that I am considerably less active here the last few days
22:27:45Araqyeah but it's always like this
22:27:51BitPuffinToday was actually the first time I did some gaming in months or so because a game I have been looking forward to a long time came out today
22:28:02BitPuffinYeah
22:28:11BitPuffinWe need some nimrodders who don't have anything to do
22:28:12BitPuffinhaha
22:28:17Araqgta ?
22:28:22BitPuffinNah
22:28:31BitPuffinAmnesia: A Machine for Pigs
22:29:07Araqmeh, I might watch a "let's play" about it during a debugging session :P
22:29:25BitPuffinMaybe you should watch mine then :P
22:38:38BitPuffinDon't remember the last time I played a game with such great audio
22:39:53BitPuffinAraq: I dunno, at the same time it is something that needs to be experienced. And maybe THEN look at a let's play and laugh at their misery
22:47:04AraqBitPuffin: make something like this in nimrod: http://www.voxelquest.com/1/post/2013/09/first-real-preview.html
22:47:57BitPuffinHey!
22:47:59BitPuffinHere's an idea
22:48:01BitPuffinyou do it ;)
22:48:06BitPuffinno but seriosly
22:48:12BitPuffinWhat I am currently making with nimrod
22:48:39BitPuffinIs to me more exciting than voxel stuff
22:49:14Araqgood
22:50:02BitPuffinAraq: Basically I am hoping to put some stuff akin to this in a multiplayer environment http://graphics.berkeley.edu/papers/Parker-RTD-2009-08/index.html
22:50:24BitPuffinWhich is basically why I have been inactive.. Been reading math stuff :P
22:51:08*shodan45 quit (Quit: Konversation terminated!)
22:51:41Araqthat's a good way to spend one's time
22:51:44Araqgood night
22:51:48BitPuffinyep
22:51:49BitPuffinnight!
23:18:18*MFlamer joined #nimrod
23:19:38MFlamerHi there. Is the difference between ptr and pointer just that ptr is a "typed" pointer and pointer is a raw "void" pointer?
23:20:07BitPuffinMFlamer: There is "pointer" ?
23:22:12BitPuffinMFlamer: Where are you seeing this? Because afaik there is only ptr
23:22:16BitPuffin(and ref)
23:23:07MFlameryes, it's a C pointer. It's what calls to alloc and allocshared return
23:23:26BitPuffinMFlamer: link?
23:23:49MFlamerhttp://nimrod-code.org/system.html#612
23:23:57MFlamersearch on alloc
23:24:21BitPuffinhmm
23:24:24BitPuffinyou seem to be right
23:24:34BitPuffinBut it looks like "pointer" is a type
23:24:49BitPuffinwhereas ptr is an erm
23:24:52BitPuffinwhat's it called
23:25:13MFlameryou can cast it to a ptr of some type. I guess I answered my own question. It;'s just a raw address
23:25:29BitPuffinMFlamer: Yeah probably, wrapped in a type
23:26:12MFlamerok, thanks
23:27:30BitPuffinMFlamer: https://github.com/Araq/Nimrod/blob/master/lib/system.nim#L43
23:28:02MFlamerHey, do you know how to use a type class to restrict generics to a range of types like int8 through int64 and float32 and float64 and ptr? Basically like any scalars or pointers?
23:28:52BitPuffinMFlamer: I suppose the answer is that you know it if it is what seems right
23:28:59BitPuffinguess it kind of depends on the situation
23:29:42MFlamerok, so maybe it's not necessary
23:30:02BitPuffinI think maybe it's when you know something wouldn't work with say an int8 but the compiler wouldn't fail the compilation because the code is still valid
23:30:04BitPuffinBut yeah
23:30:13BitPuffinI'd say don't think about that too much
23:30:18BitPuffinYou'll notice when you need it
23:32:40BitPuffinif ever :P
23:43:35BitPuffinMFlamer: I am pretty sure the feature is mostly for saying that "this proc only takes number" or "this proc only takes integer types" etc. Not so much for saying that it doesn't work with int8 or whatever, but I suppose it could happen
23:45:31*Associ8or joined #nimrod
23:45:31*Associ8or quit (Changing host)
23:45:31*Associ8or joined #nimrod
23:47:05*Associat0r quit (Ping timeout: 245 seconds)
23:47:27MFlamerWhat I'm trying to imply is that this proc can take any T as long as T is a scalar of size 8-64bits or T is a pointer. So would the type class restriction be "char|int|float|ptr|pointer"
23:48:09BitPuffinMFlamer: Hmmm nooo
23:48:37MFlameror, like you said, its so general just let the compiler complain if there is a problem
23:49:21BitPuffinMFlamer: I think at least that you need to replace the number stuff listed with TNumber
23:49:30BitPuffinMFlamer: Not so sure about the ptr|pointer part
23:49:45BitPuffinno ptr should be there probably
23:49:49BitPuffinbecause it is not a type
23:50:07BitPuffinthen you probably makes a proc that takes foo: ptr T
23:50:20BitPuffinNot 100% on this one though
23:50:31BitPuffintaking the pointer type should be possible though
23:50:50MFlamergood point on ptr
23:51:01MFlamerno pun intended
23:51:12BitPuffinhehe
23:51:33BitPuffinMFlamer: Some useful type classes https://github.com/Araq/Nimrod/blob/master/lib/system.nim#L52
23:56:54MFlamerperfect, thanks